Polyurethane Foam Wristbands as Personal Passive Samplers of Ambient Particulate Exposures to Lead and Other Metals

Abstract

Wearable wristband passive samplers are a tool for evaluating personal exposures to chemical mixtures such as semivolatile organic compounds. However, these samplers are ineffective for particulate-bound chemicals such as metals because they do not accumulate enough particles for analysis. Here, we sought to broaden the sampling to include metals by incorporating polyurethane foam (PUF) into a wearable wristband. We asked study participants (n = 82), recruited in two cohorts, to wear the wristband for five consecutive days. We also collected floor dust wipes, vacuumed dust, soil, and water from participants’ homes and collected whole blood for analysis. Metal contents (Pb, Cu, Mn, and Fe) in the PUF positively correlated with respective metal levels in dust wipes (e.g., Spearman rho = 0.54, p < 0.0001 for Pb). Pb levels in PUF wristbands and floor wipes were not strongly predictive of blood Pb levels, suggesting that participants experienced multiple routes of exposure (e.g., dust, water, diet) that were not fully captured by the PUF wristband. The levels of Pb and Cu, normalized to Fe, were lower in soil compared to the other environmental samples, indicating that particles on the PUF were similar to those in indoor dust rather than outdoor sources. Overall, these results suggest that the addition of PUF to the wristband provides a surface for particle deposition, enabling personalized assessments of ambient metal exposures with increased sensitivity relative to dust wipes.

1. Introduction

Humans are continuously exposed to a wide array of chemicals in the ambient environment, collectively contributing to an individual’s environmental exposome. Quantification of these exposures is critical to understanding how the environment affects personal health. , Semivolatile organic compounds (SVOCs) and metals are two classes of potentially toxic chemicals that can be prevalent in indoor environments, leading to concerns for exposure due to the large majority of time that most individuals spend indoors. − For instance, polybrominated diphenyl ethers (PBDEs), organophosphate esters (OPEs), phthalate plasticizers, and certain toxic metals (e.g., lead) are components of building materials and household products. − These chemical additives shed or off-gas over time and accumulate in indoor air and dust particles, resulting in exposures to individuals via inhalation, dermal contact and ingestion. High exposure to these chemicals has been associated with negative health outcomes such as carcinogenicity, reproductive toxicity, and/or neurodevelopmental toxicity. − The health impacts of low-level exposures, particularly as mixtures of chemicals, are not well known. As such, methods are needed to monitor ambient exposure to mixtures that span chemical classes and properties.

One strategy for tracking simultaneous exposure to mixtures is the use of personal, passive samplers. A notable example is polydimethylsiloxane (PDMS) silicone deployed as a wearable wristband for passive sampling of ambient SVOC exposures. SVOC concentrations on these devices have been shown to correlate significantly with biomarker concentrations − and, to a lesser extent, with environmental SVOC concentrations (e.g., dust or air). −

Wristbands offer benefits relative to those of both biomarkers and ambient environmental samples. Compared to biomarkers (e.g., blood, urine), wristbands are less costly, less invasive, and more suitable for long-term storage. Also, the concentrations of biomarkers, particularly in urine, can fluctuate due to metabolic processes and the short half-lives of some chemicals, confounding their interpretation in ways that are not a factor for wearable passive samplers. , Compared to ambient environmental samples, wearable samplers travel with the individual, allowing for a single sampler to be used in place of multiple samples (e.g., stationary air samplers, dust, wipes, soil). Furthermore, wearable devices allow for sampling of the personal dust cloud, created from the suspension of particles as the individual moves in an indoor space and also from shedding or off-gassing from the body. Passive wristbands are also less burdensome and more suitable for daily wear than active sampling pumps. ,

Silicone wristbands have primarily been used and were originally designed for gas-phase organic compounds because of the high gas permeability of PDMS. ,, However, for SVOCs associated with suspended particulates rather than the gas phase, wristband concentrations have also been shown to significantly correlate with levels in environmental and biomarker samples. ,, These findings suggest a mechanism enabling particle-bound chemicals to accumulate on wearable wristband samplers and support adaptation of the wristband sampling concept to include metals, a chemical class that occurs strictly in the form of dust and aerosol particles rather than the gas phase in indoor settings.

Although many metals can be hazardous to human health, lead (Pb) in particular remains a concern because of its widespread occurrence in legacy products (e.g., paint, leaded gasoline, water pipes) , and present-day exposures to individuals from multiple sources (e.g., dust, soil, drinking water, and food). − Blood lead levels are frequently a benchmark for Pb exposures. However, at low levels of Pb, this biomarker could represent recent exposures or Pb remobilized from the bones where accumulation occurred earlier in life. , As such, environmental Pb sampling provides a means toward identifying present-day exposure sources and potential intervention strategies. Co-occurring metals such as cadmium (Cd) that might be enriched in Pb-based products could also be part of the exposure profile for individuals vulnerable to ambient Pb exposure. , Therefore, a personal sampling device that captures exposures to a broad suite of chemicals (e.g., metals and SVOCs) would be beneficial, especially for residents vulnerable to exposures from legacy products in older homes and from modern-day materials. Unfortunately, silicone wristbands are not appropriate samplers for metals because silicone does not accumulate sufficient amounts of suspended dust and aerosol particles during deployment. In contrast, polyurethane foam is a high-specific-surface-area material that has been used to collect aerosol particles in consumer vacuum cleaners , and functions as a passive sampler for metals in atmospheric aerosols. Foam substrates have been used previously for personal passive sampling − but not for particulate-based metal exposure.

Here, we sought to modify the wearable wristband sampling approach to include metals by testing polyurethane foam (PUF) as a material that can be attached to a silicone wristband and act as a passive sampler for ambient dust and aerosol particles. The study was motivated by the need to expand the variety of chemical exposures that can be evaluated with a single wearable sampling device. A wearable device with this expanded capability could decrease the need for ambient sample collection for metals exposure analysis (such as indoor dust wipes, soil, and aerosol particles). We constructed a “foam wristband” wearable sampler comprising a precleaned PUF segment inserted into the cavity of a silicone watchband holder. Participants were recruited in two cohorts: (1) women enrolled in a longitudinal pregnancy study (n = 50); and (2) adults in a neighborhood community in Durham, North Carolina (n = 32). These participants were asked to wear a PUF wristband for 5 days. The levels of Pb and other metals were then quantified in the foam segments and compared to levels in environmental samples (e.g., floor dust wipes, vacuum dust, and soil) collected from the participants’ homes. We also examined the relationship between Pb levels in these samples and blood Pb levels of the participants. Finally, we explored the relative quantities of multiple elements (e.g., Cu, Fe, Mn, etc.) on the foam wristbands and in other environmental samples to understand the potential sources of Pb that accumulate on the foam.

2. Materials and Methods

2.1. Foam Wristband Sampler and Other Materials

Materials for the passive samplers include PUF disks (Tisch Environmental TE-1014; density = 0.022 g/cm³) and silicone watchbands marketed for electronic fitness tracker devices (Amazon.com supplier QGHXO). The watch compartment of the watchband (Figure A) has an opening (ca. 0.5 cm × 1 cm × 4 cm) that can accommodate a PUF segment. The sides of this compartment were cut out to maximize exposed surface area for the foam. Both foam and watchbands were washed overnight in 2% HCl, rinsed with deionized water, and then dried at 60 °C. Rectangular foam plugs (∼0.38 g each) were cut from the PUF disks and inserted into the pocket of the watchband (Figure B). Further details are presented in SI Text 1. Following assembly, wristbands were stored in resealable plastic bags until given to participants. Unused foam wristbands (n = 9) were stored and analyzed as blanks.

1.

Foam wristband passive sampler: (A) watchband and polyurethane foam segment; (B) assembled wristband sampler; and (C), (D), and (E) examples of worn foam segments with varying degrees of discoloration after study participants wore a wristband for 5 days.

Precleaned 50 mL polypropylene centrifuge tubes (VWR 89049-174) were used for acid cleaning sampler components and for the collection of water samples. Precleaned 150 mL wide-mouth clear glass jars were used for the collection of outdoor soil. Floor dust was collected with wipes (GhostWipe, SC4050). Wipes, vacuumed dust, and soil were processed in polypropylene digestion tubes (Environmental Express UCC000-50). Trace metal-grade or optima-grade HCl, HNO₃, and H₂O₂ (Fisher Scientific) were used for acid cleaning steps and for sample digestion. Ultrapure (>18 MΩ, Millipore Milli-Q) deionized water was used in all cleaning and analytical procedures.

2.2. Study Participants

Adult participants for this study were recruited from two cohorts. The first included participants of the HOPE1000 study, a longitudinal pregnancy cohort that recruited pregnant women from the Durham, North Carolina (NC) area between 2018 and 2024. For this work, we contacted and enrolled 50 HOPE1000 participants for a follow-up substudy involving the foam wristband and collection of environmental samples from each participant’s home.

The “Community” cohort was recruited during an outdoor social event at a Durham, NC, neighborhood and during subsequent visits to the neighborhood in October 2024. This neighborhood was selected based on the age of the homes and a prior report of lead-based paint present in the homes within this neighborhood. A total of 33 adult participants (24 female, 7 male) were enrolled for this study.

Participant enrollment and data collection were conducted according to the Duke University Medical Center Institutional Review Board (Pro00110402 and Pro00100000) for HOPE1000 and the Duke University Institutional Review Board (2025-0040) for the Community cohort. All participants provided informed consent prior to participating in the study activities.

2.3. Sample Collection

Participants in both cohorts were asked to wear a foam wristband continuously for 5 days. Participants were requested to remove the wristband during any activities where the foam could be in contact with water, such as bathing and dishwashing. The participants were asked to mark the start and end times/dates of the 5 day period and to place the wristband in the sample bag at the end of this period. Wristbands were returned to study coordinators and stored in resealable plastic bags at room temperature until they were further processed.

We also collected environmental samples from the participants’ homes (e.g., floor wipe, water, outdoor soil, and indoor dust) and whole blood from participants (Figure S1), according to the following procedures:

Floor dust wipes for metals analysis were collected in a 1 ft² area, based on protocols outlined in the USEPA “Lead Dust Sampling Technician Field Guide”. For participants of the HOPE1000 cohort, wipes were collected from three rooms in their homes, selected from participants reporting the most frequent use. In a few homes, a subset of floor wipes was collected from carpeted floors (rather than hard surface floors), which might be a source of elevated dust loadings; however, we did not observe differences in the overall trend by excluding these samples from the metals analyses. Thus, for simplicity, the metal contents of these floor wipes are reported as the average of the three wipe locations. For the Community group, a single floor dust wipe for each person was collected from the common room near the home entrance. Collected wipes were placed directly into a digestion tube after sampling.

Water samples were collected from each participant’s home and acidified to approximately 1% (v/v) HNO₃ upon return to the lab. For participants of the HOPE1000 cohort, we requested access to “drinking water”, which could include unfiltered water from the tap or filtered water. Water samples for the Community cohort correspond to unfiltered tap water from the participants’ homes.

Outdoor soil samples were collected from the homes of HOPE1000 participants only. Surface soil samples were collected with a hand trowel at three locations in each participant’s yard (near the home foundation, near the street side of the yard, and at a midpoint between the home and the street). Soil from each location was placed into a 150 mL jar. Upon return to the laboratory, the soils were oven-dried at 50 °C for at least 3 days. Dried samples were then homogenized by passing them through a 2 mm stainless-steel sieve and grinding with a mortar and pestle. Each soil sample was analyzed individually for elemental content. Here, we report the elemental content data as the average of the three yard locations for each home.

Vacuumed dust samples were collected from HOPE1000 participants only. A cellulose thimble was placed at the end of the Eureka Mighty Mite vacuum wand (model 3670G), between the head attachment and the extension tube. Dust was collected into this thimble by vacuuming the floors of the same three rooms from which a dust wipe sample was taken. Each room was vacuumed for a specific amount of time based on its area (2 s per square foot). Upon return to the laboratory, the dust was emptied from the thimble and sieved to the <150 μm fraction for chemical analysis.

Whole venous blood was collected from each participant by a trained phlebotomist and stored in EDTA-lined Vacutainer blood specimen vials (certified for trace Pb analysis). Upon sample collection, the filled tubes were inverted several times, stored on ice during transport, and frozen in the laboratory until analysis. Blood samples from the Community cohort were collected within 13 days of when the participant had stopped wearing the wristband. For HOPE1000 participants, blood was collected during one of the prenatal care visits scheduled for the parent study. This timing was prior to enrollment in this substudy, which spanned from days to several months. For this study, we limited blood analyses of HOPE1000 participants to samples collected within 45 days of wearing the foam wristband (n = 26).

2.4. Sample Processing and Elemental Analyses

All solid samples were prepared for analysis by acid digestion. All foam samples and dust wipe samples from the Community cohort were digested with an HNO₃–HCl–H₂O₂ method. First, we added 5 mL HNO₃ and 2 mL HCl and allowed the samples to reflux for 2 h at 97.5 °C in a temperature-controlled hot block (Environmental Express). Then, 1.5 mL of 30% H₂O₂ was added to each tube and heated for an additional 1.5 h. Samples were then cooled, diluted to 50 mL with deionized water, and stored at room temperature until analysis. Vacuumed dust and soil samples were processed with an HF–HNO₃–H₂O₂ digestion. 2 mL each of HF and HNO₃ were added to ∼ 50 mg of sample. The mixture was refluxed at 105 °C for 24 h, then dried and reconstituted in 1 mL HNO₃, 1 mL H₂O₂, and 5 mL of deionized water. This mixture was refluxed overnight, and then an aliquot was collected for analysis. Dust wipes for the HOPE1000 cohort were processed with the HNO₃–HCl–H₂O₂ method followed by the HF–HNO₃–H₂O₂ method. Each digestion batch, comprising up to 36 samples, included at least one of the following matrix-matched certified reference materials (CRM): (1) National Institute of Standards and Technology (NIST) 2583“Trace Elements in Indoor Dust”; (2) NIST 2710a“Montana I Soil”; (3) NIST2711a“Montana II Soil”; or (4) BCR-320R“Channel Sediment”.

Whole blood samples were processed by heated acid digestion of 0.5 mL of blood with 1 mL of concentrated HNO₃ and 0.5 mL of concentrated HCl. The mixture was heated to at least 65 °C for 1–2 h in a temperature-controlled hot block and then heated for another hour after the addition of 1 mL H₂O₂. After cooling the digestates, a 10 mL aliquot of a 4 mg/L gold + 2% HCl solution was added to aid in mercury stability in the digestates. Each digestion batch of up to 36 blood samples was accompanied by certified reference materials: (1) NIST 955d“Toxic metals and metabolites in frozen human blood” and (2) International Atomic Energy Agency (IAEA)-A-13“Trace elements in freeze-dried animal blood”. Measured levels of Pb concentration in the blood reference materials were within 90–110% of the certified values.

Aliquots of acid digestates were diluted in a solution of 2% HNO₃/0.5% HCl, spiked with internal standards (Sc, Ge, Y, Rh, In, Tb, and Bi), and then analyzed by inductively coupled plasma mass spectrometry (ICP-MS) (Agilent Technologies 7900 or Thermo Fisher X-Series II). Analysis was performed in helium kinetic energy dispersion gas mode for most target analytes; Ca and Fe analyses were performed in hydrogen reaction gas mode. The analysis of an aqueous CRM (NIST 1643f or HPS trace metals in drinking water) was performed during each ICP-MS batch run to check the quality of the instrument calibrations. External precision for most elements is typically 4% or less based on replicate analyses of samples and standards.

Target analytes included Pb and 21 other elements (Na, Mg, Al, K, Ca, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Rb, Sr, Cd, Cs, Ba, and U). We report a subset of these analytes (Pb, Cu, Mn, Fe), selected based on performance quality control measurese.g., comparability of CRMs between digestion methods (Table S1), analyte recoveries in CRMs (Table S2), and sample measurements relative to procedural blanks (Table S3). For this study, Pb was the focus because of its relevance to human health. Cu, Mn, and Fe were included in the analysis as potential tracers of particles (e.g., soil, indoor dust) and because of the observed data quality for these analytes (described further in the Results). Three other elements (Na, Ca, Zn), shown in the Supporting Information, had lower confidence in agreement between digestion methods (Na, Ca) or a high proportion of measurements near blanks (Zn in wipes, Cd in foam, and wipes).

For each target element in blood, vacuumed dust, and soil, the measured concentration in the method blank of the digestion batch (e.g., reagents digested in empty containers with no specimen) was used for blank subtraction of the respective measurement in sample digests. We defined lower limits of quantification (LOQ) for each element in blood, vacuumed dust, and soil as the mean +2 × standard deviation (SD) of the method blanks. We also analyzed “sampler blanks” for the foam and dust wipes, which entailed clean and unused PUF segments (n = 9) and dust wipes (n = 12) processed via acid digestion in the same batches as the samples. For these sampler blanks, we calculated the mean value for each analyte and the “98th percentile” value of the blank (defined as the mean +2 × SD).

2.5. Data Analysis

All values of Pb were above the LOQ for all sample types. However, for some other elements in a selected number of soil, blood, and vacuumed dust samples, the concentration values were below the LOQ. For these samples, the concentration values were replaced with a random value between 0 and the LOQ prior to statistical analysis and data visualization. For foam and dust wipes, element concentration values were compared alongside the sampler blank mean and 98th percentile value for the respective analytes.

For one study participant in the Community cohort, the levels of major elements (Al, K, and Fe) were at least 10 to 100 times lower than those in the other foam samples and were similar to the foam blanks. As such, we presumed that this participant may not have worn their foam wristband as directed and, therefore, excluded data from this participant from all analyses in this study.

Statistical analyses were performed by using Python or JMP statistical software. Some data distributions were non-normal, even after log transformation; therefore, nonparametric statistical analyses were used. Spearman correlations between Pb measured in paired sample types (e.g., foam and dust wipes, blood and water, etc.) and associated p-values were calculated using the scipy.stats package in Python. Statistical significance was defined as p < 0.05.

To assess potential sources of metals that accumulated on the foam wristbands, we normalized the contents of Pb, Cu, and Mn to Fe content (assuming Fe was a marker of soil or other geogenic particles). To test for differences in element ratios between sample types (e.g., Pb/Fe on foam vs Pb/Fe on wipes) for the HOPE1000 and Community cohorts, we used the Wilcoxon signed-rank test. Differences in the element ratios between cohorts were tested by the Wilcoxon rank-sum test.

3. Results and Discussion

3.1. Lead Levels in Floor Wipes, Dust, Soil, and Foam Wristbands

The levels of Pb in floor wipe, dust, and soil samples were within the expected range of values for North American homes (Figure ). Overall, the Pb values measured in vacuumed dust (median = 19 μg/g), soil (median = 19 μg/g), and floor wipes (median = 0.23 μg/ft²) for the HOPE1000 cohort were similar to or below levels in the general population reported in the American Healthy Homes Survey II (AHHS II). Pb levels in the wipes for the Community cohort (median = 1.9 μg/ft²) were higher than those of the HOPE1000 participants and were comparable to levels reported in a previous study with similar home ages. Between both cohorts, four dust wipes contained Pb levels that exceeded the proposed EPA lead abatement threshold of 5 μg/ft². , One vacuum dust sample and one soil sample contained Pb levels greater than the EPA lead soil screening threshold of 200 μg/g. , These six samples that exceeded benchmark values for Pb risk were from different participant homes.

2.

Pb levels on foam wristband samplers worn for 5 days by participants and Pb levels in (A) floor dust wipes; (B) vacuumed indoor dust (HOPE1000 only); and (C) outdoor soil composite samples (HOPE1000 only) collected from the homes of study participants. Each panel shows the Spearman correlation coefficient r _s and p-value of data from the HOPE1000 cohort (blue text) or combined data of both cohorts (black). The dotted lines in panel (A) refer to the mean and 98th percentile of Pb content in clean, unused wipes.

For PUF wristbands, the median Pb value across both cohorts was 0.14 μg/g PUF, or 8.6 × 10^–3 μg per cm². (The exposure area of each PUF segment was about 6 cm² per segment, and each PUF segment was approximately 0.38 g). We are not aware of a previous study that has measured metals in a wearable PUF sampler. However, in a prior use of PUF as a stationary passive sampler of outdoor aerosols, the PUF accumulated approximately 0.01 μg-Pb per g PUF after 1 week of sampling, corresponding to 2 × 10^–4 μg per cm². Thus, the Pb values observed in the foam for our study are at least 10 times greater than those observed through stationary aerosol sampling. Notably, outdoor stationary samplers are designed for diffusive-based deposition of particles and exclude settled particles. In contrast, the wristband format does not control particle deposition pathways, including diffusive uptake, settling, and direct contact with dust-laden surfaces. Differences in the sampling environment (i.e., indoor vs outdoor) could also explain the discrepancy in the Pb loading values in this comparison. Moreover, increased particle loadings on the wearable foam sampler might also be due to a personal cloud effect, caused by particles shedding from the individual or resuspended from surfaces as the individual moves through their environment. ,

When comparing Pb levels on the PUF with levels in other environmental samples, we observed PUF Pb values to significantly correlate with Pb levels in dust wipes (r _s = 0.54, p < 0.0001) (Figure ). We also observed a statistically significant, weaker correlation between PUF Pb levels and soil Pb levels (r _s = 0.31, p < 0.05). Notably, Pb contents in a considerable number of dust wipe samples were less than the 98th percentile for blank wipes (Table S3; Figure A, dotted lines), indicating that dust wipes may not reliably assess metals exposure at low levels. Previous studies have similarly found that many values were below method detection limits for the use of dust wipes to monitor Pb levels in homes within Canada and the United States. − Conversely, Pb PUF levels were above the blanks for all participants except for one with a Pb value (1.1 × 10^–3 μg per cm²) that was less than the blank 98th percentile (1.2 × 10^–3 μg per cm²). This result suggests that PUF has a lower relative limit of quantitation for Pb, which is an advantage over floor dust wipes, a widely employed method for environmental Pb exposure assessments. ,,−

The Community cohort, in general, had higher levels of Pb in both the foam and dust wipes than the HOPE1000 cohort (Figure A). This finding was expected given that the median year of building construction in the Community cohort (1954) was older than the HOPE1000 cohort (1990–1999) and predated the phase-out of Pb-based paint in the U.S. in 1978. Indeed, we observed higher levels of Pb in most samples for older homes (Figure S2). These results are in agreement with previous studies, in which home age has been documented as a predictor of Pb levels in house dust ,, and soil, such that older homes are likely to be associated with higher Pb levels. Additionally, higher Pb wipe values in the Community cohort may be partly attributable to sampling near entryways, which have been previously documented to contain greater metal loadings than interior locations. , Still, this effect has been measured to be less than the order of magnitude difference between the two cohorts observed here. Overall, the PUF wristband’s ability to detect higher exposure in the Community cohort compared to the HOPE1000 cohort supports its effectiveness in detecting differences in environmental metal exposure.

For other metals such as Cu, Mn, Fe, Ca, Na, Zn, and Cd in foam wristbands and environmental samples, we observed many of the same correlation trends as those for Pb (Figures S3–S9). Notable exceptions were the levels of Mn and Fe in foam, which did not correlate with their respective levels in soil samples (Figures S4 and S5).

3.2. Foam Wristbands and Environmental Samples as Indicators for Internal Pb Dosage

Pb levels in foam wristbands and dust wipes were poorly correlated to blood Pb levels for participants in both cohorts (Figure A,B). This analysis of blood data was performed separately for each cohort rather than combined across cohorts because of key differences between the groups that prevent integration of the data sets. For example, the participants of HOPE1000 were all pregnant women. Blood volumes and Pb concentrations can vary appreciably during pregnancy, resulting in blood data for the HOPE1000 cohort that may not be comparable to those in the Community cohort.

3.

Blood Pb levels in study participants and Pb levels in (A) foam after participants wore the wristbands for 5 days; (B) floor dust wipes from the participant homes; and (C) “drinking water” for the HOPE1000 cohort or tap water for the Community cohort. Spearman correlation coefficients r _s and p-values refer to data of the HOPE1000 cohort (blue text) or the Community cohort (orange).

For the HOPE1000 cohort, blood Pb levels were generally lowthe highest blood Pb level for this cohort was 5.9 μg/L, a value that is less than the median level for U.S. adults in the 2018 NHANES survey (8.5 μg/L). This result suggests that Pb exposures among these study participants were low for all potential exposure pathways (e.g., dust, aerosols, water, diet). At these levels, exposures to Pb are likely to occur through multiple pathways, including ambient dust, aerosols, water, and diet; thus, differentiation between these exposures was not possible. The blood Pb data for HOPE1000 participants are further confounded by the blood collection time, that was up to 10 weeks prior to the collection of foam, wipes, and water samples (Figure S1). Thus, pairing of the blood samples with the environmental samples may not be appropriate if study participants experienced large variations in exposure rates during the ` period between blood collection and home sampling.

For the Community cohort, blood Pb had a weak positive correlation with Pb levels in the foam wristbands (r _s = 0.13) and dust wipes (r _s = 0.20), but correlations were not statistically significant (Figure A,B). The correlation between blood Pb and water Pb was moderate, yet statistically significant (r _s = 0.38, p < 0.05) (Figure C). Thus, the data indicates that tap water could be an important contributor to exposure. We note that all tap water samples from these homes were below the EPA’s proposed action level of 10 μg/L. The median blood Pb measured in the Community cohort (7.67 μg/L) was lower than the median blood Pb for adults in NHANES (8.5 μg/L). However, six participants had blood Pb above the 95th percentile (26.2 μg/L) reported in the NHANES data set. For these individuals in particular, there may be additional routes of exposure, such as diet, contributing to their blood Pb, given that none of their water measurements were above the EPA action level. Moreover, blood Pb analyses could represent an exposure time frame that is longer than the 1 week that the participants were involved in our study. For example, blood Pb levels can also be affected by internal mobilization of Pb from bones, which represent a reservoir of long-term exposure. ,,

3.3. Sources of Particles Accumulated on Foam Wristbands

The accumulation of Pb and other metals on the foam samples could originate from multiple sources such as ambient house dust and soil particles from the participant’s environment. To differentiate potential sources of particles, we examined the levels of Pb, Cu, and Mn in relation to Fe (Figure ). Fe was selected as a reference element due to its high crustal abundance and because extraction efficiencies between digestion methods were comparable (i.e., within 35% relative difference) for all certified reference materials (e.g., dust, sediment, and soil) (Table S1). In addition to Pb, we examined the relative levels of Cu and Mn, elements that are potential representations of house dust and soil-based particle sources , and were chosen because their levels in most sample wipes were greater than the blanks (Figures S3 and S4).

4.

Relative ratios of (A) Pb:Fe; (B) Cu:Fe; and (C) Mn:Fe in worn foam, floor dust wipes, vacuumed dust, and outdoor soil collected from individuals in the HOPE1000 and Community cohorts. Box plots depict the 25th to 75th percentiles (box outline), median (horizontal line), and 5th to 95th percentiles (whiskers) of sample points (all plots in circles) within each group.

Overall, we consistently observed that the element ratios on foam samples were within the same range of values as those ratios for floor wipes and vacuumed dust, while element ratios in soil were different from those in the other sample types (Figures , S10). For example, the ratios of Pb:Fe and Cu:Fe in foam, wipes, and dust (Figure A,B) are significantly elevated above soil (p < 0.0001). These findings are corroborated by other elements (Na, Ca, Zn, Cd) when normalized to Fe (Figure S10). Mn:Fe ratios were an exception to this observation (Figure C): all Mn:Fe values for the 4 types of samples were within ∼1 order of magnitude of each other. No significant difference was observed between Mn:Fe in foam and Mn:Fe in soil (p = 0.19). Together, these results suggest that the source for Mn in these samples is the same as Fe, likely outdoor soil.

We also found that when Pb, Cu, and Mn levels were normalized to Fe (Figure ), differences between cohorts diminished, indicating that cohort variation reflects total particle loading rather than differences in the metal composition. This pattern is consistent with prior findings that, in the general population, particle mass loading is a stronger predictor of metal exposure than metal concentrations in ambient particles.

3.4. Environmental Implications

Overall, this study demonstrates the utility and potential limitations of a wearable wristband modified with a polyurethane foam surface for personal metal exposure assessment. The foam wristbands clearly accumulated particles from each individual’s environment – the levels of metals contained in the foam after a 5 day period were above blanks for the majority of participants in this study. Metal levels on the foam were also shown to correlate with their levels in dust wipes collected from the participants’ homes. Because of the low blanks enabled by precleaning the PUF, the foam wristband approach had improved sensitivity for Pb exposure assessments compared to dust wipes, which had higher blanks for Pb.

Because of the design of the wristband, particles could accumulate on the foam through direct contact with dusty surfaces or through deposition of aerosolized particles from the individual’s immediate surroundings. These processes are expected to have variable rates of particle accumulation, resulting in concentrations of metals on the foam that cannot be used to precisely model or extrapolate to metals concentrations in airborne or settled particles. Regardless, the levels of metals such as Pb on foam are approximations of ambient exposure levels, suggesting that a relatively simple modification to the wearable wristband could broaden exposure assessments of environmental mixtures. While we do not propose that this modified wristband would completely replace detailed ambient environmental sampling for metals (e.g., indoor dust, air, wipes), the wristband approach would be useful in circumstances where resources for environmental sampling are limited or preliminary evaluations are needed to help prioritize follow-on sampling and intervention efforts. This modified foam wristband design would also support a nontargeted approach for environmental exposure and health research.

While these findings highlight the promise of foam wristbands for environmental exposure monitoring, further application of this approach will need to address several challenges and limitations. In our study, we could not independently verify compliance among participants who were instructed to wear the bands for 5 days. In some foam specimens, particles were not visible, or we measured low concentrations of all elements in our analysis panel, possibly due to incomplete or inconsistent wear by the participants. We also recognize that a wristband with a foam segment may be less comfortable to wear than the traditional silicone bands, which may reduce participant adherence to instructions. Future studies could pair wristbands with accelerometers to provide data on the participant’s compliance. Other placements of the sampler on the body could also be tested, such as the upper torso, to target inhalation exposure more exclusively. Such a repositioning would support further investigation of pulmonary exposure and health outcomes due to the inhalation of particulate matter and associated metals.

Beyond these challenges of deployment, future research is needed to determine how well metal levels on foam wristbands reflect the internal dose. The wristbands could distinguish differences in blood Pb at a broad level (i.e., across the two cohorts), but they were less reliable at the individual level. Blood Pb levels are employed by medical practitioners to understand patient health vulnerabilities from Pb exposure. However, such measures, particularly at low Pb levels, cannot readily differentiate early life from current exposures. Thus, additional sampling from the individual’s exposure environment is needed to establish intervention options. The environments occupied by many participants of this study appeared to contain low levels of Pb; hence, other exposure sources such as diet and water consumption could be more important than ambient dust and aerosols for contributing to overall Pb exposures. Additionally, exposures to Pb from ambient particles tend to occur through hand-to-mouth contact, and it is unclear whether the wristbands are capturing this uptake pathway. Future testing of the foam wristband design could focus on subpopulations with greater exposure levels (e.g., occupational settings) or with greater levels of dust ingestion from hand-to-mouth contact (e.g., children). Studies that more thoroughly investigate the uptake mechanisms and sources of Pb on the wristband may also prove useful in this regard (e.g., comparing metal profiles of wristbands to those of indoor particulate measurements, personal sampling pumps, house dust, and hand wipes). The analysis of Pb speciation on the foam wristbands could provide further insight into exposure sources as well as bioaccessibility to inhaled or ingested particles. However, the analysis will be challenging with current spectroscopic Pb speciation methods due to the relatively low levels of Pb on the foam (∼0.14 μg/g PUF).

Ultimately, the incorporation of polyurethane foam material on silicone wristbands can be a relatively straightforward modification of existing tools and commercially available materials for personalized exposure sampling. This approach would enable the simultaneous monitoring of particulate-bound chemicals with volatile chemicals in a single wearable device. With knowledge of a broad set of chemicals in an individual’s ambient environment, we can deepen our understanding of a person’s multifaceted exposure profile and potential impacts of these exposures on personal health.

Data for this article are available at the Duke University Research Data Repository. DOI: 10.7924/r4r527

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.est.6c05099.

• Additional text describing wristband preparation; tables demonstrating CRM measurements between methods and compared to certified values; a table displaying levels of metal analytes in blank foam and blank wipes; descriptive figure of sampling design; figure of Pb levels in samples grouped according to construction year of participant homes; figures showing correlations between wristbands and environmental samples for additional elements; a figure showing foam metal accumulation consistency over time; and a figure showing additional element/Fe ratios (PDF)

The authors declare no competing financial interest.